We propose to advance the study of peer influence on substance use through a longitudinal investigation of adolescent friendship networks from 6th through 9th grades. Our approach combines the strengths of the fields of research on adolescent problem behavior, developmental psychology, and network analysis. The project's specific aims are 1) to examine the developing role of substance use as a basis of similarity within friendship groups;2) to determine both the influence of friendship networks on substance use and the contribution of substance use to friendship selection;3) to determine the moderators of friends'influence on substance use;4) to examine network sources of peer influence other than the immediate friendship group, and 5) to establish the effects of an evidence-based prevention program on friendship networks and the role of those effects as mediators of the program's impact on substance use. This proposal seeks funding to analyze valuable data being collected in a large prevention trial, the specific aims of which do not include processing and analysis of the key information for our work. That study gathered five waves of data on both friendship choices and a variety of self-report measures for all students in two grade cohorts at 28 schools (N=11,000+) as well as measures of school and community characteristics. This dataset is superior to prior studies of peer influence in that it measures friends'characteristics from their own reports, includes several longitudinal assessments of both friendship networks and outcomes, covers a developmental period that is critical for both substance use and peer relations, uses rural settings that provide maximum coverage of friendship networks, and samples enough schools to study variation in peer processes across settings. We also will overcome some limitations and biases of prior research on peer influence through longitudinal diffusion analyses that capture the extended network implications of influence and friendship selection processes. Both theorists and researchers attribute great importance to friends'influence on substance use, and accordingly prevention programs emphasize peer resistance skills and use peer leaders to help convey their message. The proposed study strives to overcome the limitations in our understanding of peer influence that hamper these efforts and thereby seeks to help reduce future substance use and abuse.